Several million Americans receive long-term opioid therapy for chronic non-malignant pain. Little is known about the determinants and risks of this treatment regimen. This research focuses on back pain as a tracer condition. Health services research and epidemiologic methods will be used to understand who receives long-term opioid therapy and to assess associated risks including onset of opioid abuse and mortality. This research will be conducted in two large, non-profit health plans serving over 1% of the U.S. population. Research goals - Our goals are to advance knowledge concerning determinants and risks of long-term opioid therapy among chronic pain patients. Aim 1-Examine trends: We will describe trends in use of long-term opioid therapy in the participating health plans from 1995-2006 and determine how patient characteristics indicating clinical complexity and prior substance abuse are related to trends in long-term opioid use. Aim 2-Assess risks of opioid abuse: We will assess risk factors for opioid abuse among persons receiving long-term opioid therapy. Primary hypotheses are: 1) Among persons receiving long-term opioid therapy for back pain, use of long-acting opioids is associated with reduced risks of onset of opioid abuse and misuse relative to long-term use of short-acting opioids. 2) A prior history of drug abuse will increase risks of opioid abuse and misuse among those receiving long-term opioid therapy for back pain. These hypotheses will be tested through a case-control study comparing recent onset cases to controls (N=600). Aim 3-Assess mortality and other adverse event risks: We will assess whether mortality and other adverse event risks are increased by long-term opioid therapy. Primary hypotheses are: 1) Long-term use of opioids for back pain will be associated with increased mortality. 2) Long-term use of long-acting opioids will increase mortality risk relative to long-term use of short-acting opioids. 3) A prior history of substance abuse will be associated with increased mortality among persons receiving long-term opioid therapy. We will evaluate these hypotheses through a case-cohort study (N=1600). We will extend these analyses by studying the effects of long-term opioid therapy on other adverse events including: all-cause and opioid-related hospitalization;opioid overdose;and onset of recognized mood disorder.